Differential modulation of microglia superoxide anion and thromboxane B2 generation by the marine manzamines

Background Thromboxane B2 (TXB2) and superoxide anion (O2-) are neuroinflammatory mediators that appear to be involved in the pathogenesis of several neurodegenerative diseases. Because activated-microglia are the main source of TXB2 and O2- in these disorders, modulation of their synthesis has been hypothesized as a potential therapeutic approach for neuroinflammatory disorders. Marine natural products have become a source of novel agents that modulate eicosanoids and O2- generation from activated murine and human leukocytes. With the exception of manzamine C, all other manzamines tested are characterized by a complex pentacyclic diamine linked to C-1 of the β-carboline moiety. These marine-derived alkaloids have been reported to possess a diverse range of bioactivities including anticancer, immunostimulatory, insecticidal, antibacterial, antimalarial and antituberculosis activities. The purpose of this investigation was to conduct a structure-activity relationship study with manzamines (MZ) A, B, C, D, E and F on agonist-stimulated release of TXB2 and O2- from E. coli LPS-activated rat neonatal microglia in vitro. Results The manzamines differentially attenuated PMA (phorbol 12-myristate 13-acetate)-stimulated TXB2 generation in the following order of decreasing potency: MZA (IC50 <0.016 μM) >MZD (IC50 = 0.23 μM) >MZB (IC50 = 1.6 μM) >MZC (IC50 = 2.98 μM) >MZE and F (IC50 >10 μM). In contrast, there was less effect on OPZ (opsonized zymosan)-stimulated TXB2 generation: MZB (IC50 = 1.44 μM) >MZA (IC50 = 3.16 μM) >MZC (IC50 = 3.34 μM) >MZD, MZE and MZF (IC50 >10 μM). Similarly, PMA-stimulated O2- generation was affected differentially as follows: MZD (apparent IC50<0.1 μM) >MZA (IC50 = 0.1 μM) >MZB (IC50 = 3.16 μM) >MZC (IC50 = 3.43 μM) >MZE and MZF (IC50 >10 μM). In contrast, OPZ-stimulated O2- generation was minimally affected: MZB (IC50 = 4.17 μM) >MZC (IC50 = 9.3 μM) >MZA, MZD, MZE and MZF (IC50 > 10 μM). From the structure-activity relationship perspective, contributing factors to the observed differential bioactivity on TXB2 and O2- generation are the solubility or ionic forms of MZA and D as well as changes such as saturation or oxidation of the β carboline or 8-membered amine ring. In contrast, the fused 13-membered macrocyclic and isoquinoline ring system, and any substitutions in these rings would not appear to be factors contributing to bioactivity. Conclusion To our knowledge, this is the first experimental study that demonstrates that MZA, at in vitro concentrations that are non toxic to E. coli LPS-activated rat neonatal microglia, potently modulates PMA-stimulated TXB2 and O2- generation. MZA may thus be a lead candidate for the development of novel therapeutic agents for the modulation of TXB2 and O2- release in neuroinflammatory diseases. Marine natural products provide a novel and rich source of chemical diversity that can contribute to the design and development of new and potentially useful anti-inflammatory agents to treat neurodegenerative diseases.

the fused 13-membered macrocyclic and isoquinoline ring system, and any substitutions in these rings would not appear to be factors contributing to bioactivity.
Conclusion: To our knowledge, this is the first experimental study that demonstrates that MZA, at in vitro concentrations that are non toxic to E. coli LPS-activated rat neonatal microglia, potently modulates PMA-stimulated TXB 2 and O 2generation. MZA may thus be a lead candidate for the development of novel therapeutic agents for the modulation of TXB 2 and O 2 release in neuroinflammatory diseases. Marine natural products provide a novel and rich source of chemical diversity that can contribute to the design and development of new and potentially useful antiinflammatory agents to treat neurodegenerative diseases.

Background
The hallmark of brain inflammation is the activation of glia, particularly microglia, the resident immune cells of the brain [1]. Microglia activation in brain pathologies, as caused by infectious diseases, inflammation, trauma, brain tumors, ischemia and AIDS, may result in neuronal injury and ultimately neurodegeneration [1]. Similar to other tissue macrophages, when microglia become activated they release potentially neurotoxic mediators [2], followed by sublethal and lethal injury to the central nervous system. The two different phenotypic forms of microglia, namely the activated but nonphagocytic microglia found in inflammatory pathologies and the reactive or phagocytic microglia present in trauma, infection and neuronal degeneration, appear to have the capacity to express cell-surface receptors and release mediators of inflammation, such as cytokines, coagulation factors, complement factors, proteases, nitric oxide, eicosanoids and reactive oxygen species [2].
Over the last three decades, the marine environment has been demonstrated to be a source of novel therapeutic agents, many of which have anti-inflammatory properties [3]. We have previously shown that selected marine natural products modulate eicosanoids [4,5] and O 2generation from activated rat [6] and human neutrophils [7], as well as liver [8] and alveolar macrophages [9]. Based on these observations we hypothesized that selected marine natural products might potentially attenuate activated brain microglia [2]. Since the discovery by Sakai and Higa that the marine sponge-derived manzamine A (MZA) had potent antitumor activity [10], there has been a sustained interest in the chemistry [11] as well as the pharmacology of the manzamines, a class of β-carboline marine-derived alkaloids. More than 40 manzamine-type alkaloids have been isolated from 9 different genera of marine sponges from the Indian and Pacific Oceans, and in addition to the antitumor activity [10], manzamines have been shown to be immunostimulatory [12], insecticidal [13], antibacterial [13], antimalarial [14], antiparasitic [15], antiviral [16] and to possess antituberculosis activity [17].
As shown in Fig. 2A and 3A, the cytotoxicity of MZA to neonatal brain microglia measured as LDH release was not significantly different from controls even at 10 µM (MZA vs. vehicle, respectively 21.3 ± 7 % vs. 13.9 ± 3.7 % of total LDH released by 0.1 % Triton X-100 treated- The chemical structures of manzamines A, B, C, D, E and F Figure 1 The chemical structures of manzamines A, B, C, D, E and F. Manzamines are indole-derived alkaloids isolated from the marine sponges Haliclona sp. [10], Amphimedon sp. [66] and Xestospongia sp. [10,67]  Effect of manzamine B on LPS-activated neonatal brain microglia TXB 2 , O 2 and LDH release MZB differs from MZA in having a tetracyclic diamine complex and an epoxide ring system (Fig. 1). As shown in Fig. 2B, MZB which was tested as a free base, was less potent than MZA in affecting PMA-stimulated TXB 2 generation (IC 50  As shown in Fig. 2B and 3B, in contrast to MZA, MZB was cytotoxic to neonatal brain microglia at concentrations above 1 µM. In fact, considerable LDH release was observed at 10 µM (88.3 ± 12 % of total LDH released by 0.1 % Triton X-100 treated-microglia, n = 5, P < 0.01). Taken together, these data suggest that the reduction of both O 2and TXB 2 generation resulted from both pharmacological and toxic effects of MZB on LPS-activated microglia cells.

Effect of manzamine C on LPS-activated neonatal brain microglia TXB 2 , O 2
and LDH release MZC differs from MZA in having a monocyclic amine ring attached to the β-carboline moiety (Fig. 1). As shown in Fig. 2C, MZC which was tested as a free base, was less potent than MZA in affecting PMA-stimulated TXB 2 generation (IC 50  As shown in Fig. 2C and 3C, and in contrast to MZA, MZC was cytotoxic to neonatal brain microglia though not as much as MZB. Substantial LDH release was observed at 10 µM (59.8 ± 11 % of total LDH released by 0.1 % Triton X-100 treated-microglia, n = 5, P < 0.05). In summary, similar to MZB, the data suggests that the reduction of both O 2and TXB 2 generation resulted from both pharmacological and toxic effects of MZC on LPS-activated microglia cells.
As shown in Fig. 2D and 3D, and in contrast to MZB and C, MZD was very cytotoxic to microglia when PMA was used as an agonist to trigger O 2and TXB 2 release: MZD at 0.1 µM caused 61.5 ± 13 % of total LDH released by 0.1 % Triton X-100 treated-microglia (n = 5, P < 0.05). In contrast to the limited effect of MZD on OPZ-stimulated microglia, the data suggests that the reduction of PMA-stimulated O 2and TXB 2 generation resulted from both pharmacological and toxic effects of MZD on LPS-activated microglia cells. As shown in Fig 2E and 3E, cytotoxicity of MZE to microglia measured as LDH release was low even at 10 µM (MZE vs. vehicle, respectively 20 ± 6.7 % vs. 19.8 ± 7.7 % of total LDH released by 0.1 % Triton X-100 treatedmicroglia, P > 0.05, n = 5).

Effect of manzamine F on LPS-activated neonatal brain microglia TXB 2 , O 2
and LDH release MZF differs from MZA in having a saturated ketone functionality in the eight-membered amine ring and hydroxylation at the C-8 position of the β-carboline ring system ( As shown in Fig. 2F, MZF effect on OPZ-stimulated TXB 2 generation was weak, with a non-statistically significant 26.2 % inhibition at 10 µM (MZF vs. vehicle, respectively 3,317 ± 121 vs. 4,504 ± 308 pg of TXB 2 per 200,000 microglia per 70 min, P > 0.05, n = 2). Similarly, as depicted in Fig. 3F, MZF was minimally effective in inhibiting OPZ-stimulated O 2generation, only a non-statistically significant 15.7 % inhibition observed at 10 µM.

Effect of manzamine A, B, C, D, E and F on hypoxanthinexanthine oxidase generated O 2 -
In order to determine a potential scavenging effect of MZA, B, C, D, E and F on O 2 -, a standard hypoxanthinexanthine oxidase system was used as a cell-free source of O 2 - [21]. As shown in Fig. 4

Discussion
The important role of neuroinflammation and glial activation in the pathogenesis of brain disorders has progressively been established [1,2,22]. Because in vitro LPSactivated microglia appear to mimic the functions of activated microglia found in neuroinflammatory conditions in vivo [2,23], we used LPS-activated rat microglia as a relevant in vitro paradigm to search for marine natural products that may modulate the enhanced release of TXB 2 and O 2from activated microglia [24]. Using this in vitro model we have previously communicated that MZA, a secondary metabolite isolated from the Okinawan marine sponge Haliclona sp [10], inhibited TXB 2 and O 2generation by microglia [18]. The current study extends our initial observations, and reports a structure-activity relationship study with manzamines A, B, C, D, E and F on both PMA and OPZ-stimulated release of TXB 2 and O 2 from LPS-activated rat neonatal microglia.
Members of the eicosanoid family (i.e. prostaglandins, leukotrienes and thromboxanes) are important mediators of inflammation that would appear to be play a causative role in the pathogenesis of several CNS disorders [25][26][27].
Increased levels of eicosanoids have been observed in neurodegenerative disorders such as amyotrophic lateral sclerosis [28], multiple sclerosis [29], ischemia and seizures [30], prion diseases [31], human immunodeficiency virus-associated dementia [32] and Alzheimer's disease [33]. Following the seminal observation that microglia release eicosanoids [34], numerous studies have increasingly supported the notion that activated brain macrophages may be the main source of both prostaglandins [33,[35][36][37][38][39] and thromboxanes [23,37,39,40] in these neurodegenerative diseases. Thus modulation of microglia enhanced prostanoid synthesis has been investigated as a potential drug therapeutic approach for intervention in neuroinflammatory disorders of the CNS [36,41,42]. One possible approach to diminish enhanced eicosanoid production has been to search for inhibitors of signal transduction pathways involved in eicosanoid synthesis in activated microglia [42]. In our study, we used PMA and OPZ, agonists known to activate p44/42 and p38 mitogen-activated protein kinases in microglia [43], to target distinct signal transduction pathways that cause TXB 2 release in rat neonatal microglia activated by an in vitro exposure to 0.3 ng/mL of LPS for 17 hours [23]. As shown in Fig. 2, the 6 marine-derived manzamine analogs attenuated PMA-stimulated TXB 2 generation differentially and in the following order of decreasing potency: MZA>MZD>MZB >MZC>MZE and F. In contrast, the manzamine analogs inhibited OPZ-stimulated TXB 2 generation with reduced potency: MZB>MZA>MZC>MZD, MZE and MZF. Thus, with the exception of MZB and MZC which modulated both PMA and OPZ-stimulated TXB 2 release with similar potency, MZA, MZD, MZE and MZF inhibited TXB 2 release with lower potency when OPZ was used as an agonist.
It is interesting to compare our differential results with the MZA, B, C, D, E, and F with those reported for other agents that have been shown to modulate microglia eicosanoid release by targeting the cyclooxygenase I and II enzymes which are expressed in activated rat and human microglia [35,44]. PGE 2 and TXB 2 synthesis that occurs concomitantly with LPS-induced activation of rat [38,45,46] and human microglia [35], has been shown to be attenuated by nonsteroidal anti-inflammatory drugs (NSAIDs) with differing activities towards the two isoforms of COX. Thus, LPS-induced microglia PGE 2 synthesis was reduced by COX-1 inhibitors: acetylsalicylic acid (aspirin) (IC 50 = 3.12-10 µM) [38,46], flurbiprofen (apparent IC 50 = 100 nM) [45] and indomethacin (apparent IC 50 = 1 nM) [38], and the COX-2 inhibitor NS-398 (apparent IC 50 = 1-5 nM) [38]. Even though NSAIDs have been reported to attenuate neurotoxicity in vitro [47] and neuroinflammation in animal models [48][49][50], an important caveat is the fact that determining the best NSAIDs for clinical neurodegenerative disease management appears to remain a matter of considerable debate in view of their well known adverse effects [51][52][53]. Thus, although the molecular mechanism by which the manzamines inhibit TXB 2 release in LPS-activated cells remains currently undetermined, MZA inhibited PMA-stimulated eicosanoid generation in vitro with potency similar to that of the COX-1 inhibitor indomethacin [38], potency that was higher than that of other NSAIDs that have been reported to modulate enhanced eicosanoid release in both activated rat and human microglia [35,38,45,46].
The involvement of reactive oxygen species (ROS) has been documented in CNS pathologies, such as Parkinson's disease, Alzheimer's disease, Huntington's disease, Down's syndrome, cerebral ischemia and reperfusion, amyotrophic lateral sclerosis, multiple sclerosis and meningitis [54]. Prolonged exposure to ROS may potentially damage neurons, particularly their synapses [55] as well as oligodendrocytes, the myelin producing cell of the CNS [56] by overriding normal CNS antioxidant defense mechanisms, e.g. superoxide dismutase, catalase, glutathione-S-transferase, glutathione peroxidase, permanently affecting cellular function [57]. Thus the mechanism of ROS generation by CNS leukocytes, i.e. infiltrating neutrophils and monocytes as well as resident microglia production of O 2 -, hydrogen peroxide and nitric oxide in CNS, has received considerable attention since the mid-   [2]. In fact, during the past 18 years, numerous research groups have shown that O 2may be generated by microglia isolated from rat, mice, hamsters, dogs, swine and humans, when stimulated with a variety of agonists such as phorbol ester, opsonized zymosan, calcium ionophore, antiviral antibodies, antibody-coated red blood cells and myelin (reviewed in [2]). We and others have hypothesized that rather than scavenge ROS with antioxidants, the modulation of the signal transduction mechanism leading to microglia ROS generation might be putatively a better therapeutic strategy to turn off or reduce ROS generation that could lead to neuronal injury [2,58,59]. As depicted in Fig. 3 It is interesting to compare our differential results with the MZD, A, B and C with those reported for other agents that modulate PMA-stimulated O 2 generation in microglia. Interestingly, MZD, A, B and C demonstrated higher potency than three clinically available agents shown to inhibit PMA-stimulated O 2generation: propentofylline, a selective phosphodiesterase inhibitor (IC 50 >100 µM) [60], cabergoline, a potent and selective agonist of D2dopamine receptors (IC 50 >100 µM) [59], and nicergoline, an ergoline derivative used for cerebrovascular diseases (IC 50 = 10-15 µM) [58]. It is noteworthy that these agents have been proposed to confer protective effects against neurodegenerative diseases which may involve O 2release by activated rat microglia.
In order to determine if the effect of the manzamines on PMA or OPZ-triggered TXB 2 and O 2 generation was either pharmacological or toxic, we investigated LDH release from LPS-activated rat neonatal microglia. LDH has extensively been used as a marker for cell cytotoxicity [61]. The results from our investigation appear to demonstrate that the manzamine analogs clearly differed in their effect on LDH release from LPS-primed microglia: MZA, MZE and MZF generated less than 50% of maximal LDH release at 10 µM, and thus were the least toxic analogs; MZB and MZC induced greater than 50% of LDH release at 10 µM; while MZD showed greater than 50% of LDH release at 0.1 µM when PMA rather than OPZ was used as an agonist, thus its toxicity contrasted with the other analogs tested. Thus even though MZD inhibited PMA-stimulated O 2generation with slightly higher potency than MZA, because MZD caused concomitant LDH release at low concentrations (0.1 µM), the nature of the inhibitory effect on PMA-triggered O 2 and TXB 2 release, either toxic or pharmacological, remains currently unresolved. In summary, the in vitro studies described herein suggest that MZA is the most potent manzamine analog of the series investigated because both PMA-stimulated O 2and TXB 2 were potently inhibited with the lowest concomitant release of LDH.
It is of interest to consider the results of our structureactivity relationship (SAR) study with the manzamines, alkaloids characterized by a complex heterocyclic ring system attached to the C-1 of the β-carboline moiety. From the SAR perspective, the potent effect of MZA and D hydrochloride salts on PMA-stimulated O 2and TXB 2 suggests that the solubility or ionic forms are contributing factors to their bioactivity. Furthermore, the fused 13membered macrocyclic and octahydroisoquinoline ring system, and any substitutions in these rings would appear to be less important for their in vitro activity. Finally, changes such as saturation or oxidation of the β carboline or the 8-membered amine ring tended to decrease bioactivity in both O 2and TXB 2 assays.
Taken together, our current data demonstrates that the most potent and least toxic manzamine analog, namely MZA, was less effective in attenuating O 2 and TXB 2 from LPS-activated microglia when the triggering agonist was OPZ rather than PMA. Similar differential effects between PMA and OPZ-triggered signaling have been observed with other natural products [62]. Furthermore, the current data suggest the following on the as yet undefined mechanism of action of MZA: Firstly, that the MZA molecular target plays a critical role in O 2and TXB 2 generation initiated by PMA upon binding to PKC [63,64] and activation of the p44/42 mitogen-activated protein kinase signaling pathway [43]; Secondly, that the MZA molecular target probably plays a less critical role in O 2and TXB 2 release elicited by OPZ, a ligand of the microglial cell surface complement receptor 3 shown to activate the p38 mitogen-activated protein kinase signaling pathway [43]. Studies to determine which element is targeted by MZA in the p38 and/or p44/42 mitogen-activated protein kinase pathways in LPS-activated rat microglia are currently underway in our laboratory.

Conclusion
Our present results provide the first experimental evidence to support the hypothesis that the marine-derived β-carboline alkaloid manzamines differentially modulate both O 2 and TXB 2 generated by E. coli LPS-activated rat neonatal microglia. Additional conclusions are the following: Firstly, SAR studies demonstrated that at in vitro concentrations that were non-toxic to E. coli LPS-activated rat neonatal microglia, MZA was the most potent inhibitor of O 2 and TXB 2. Secondly, although the mechanism by which MZA inhibited PMA-stimulated TXB 2 generation in vitro is as yet unclear, its potency was similar to the COX-1 inhibitor indomethacin [38], and thus higher than other NSAIDs reported to modulate enhanced eicosanoid release in both activated rat and human microglia [35,38,45,46]. Thirdly, although the mechanism by which MZA inhibited PMA-stimulated O 2 generation in vitro remains undetermined, MZA was more potent than propentofylline, a selective phosphodiesterase inhibitor, cabergoline, a potent and selective agonist of D2dopamine receptors and nicergoline, an ergoline derivative used for cerebrovascular diseases, compounds which have been proposed to confer protective effects against neurodegenerative diseases by affecting O 2release by activated rat microglia. Fourthly, SAR studies which demonstrated that the ionic forms are a contributing factor to the bioactivity of the complex manzamine heterocyclic ring system attached to a β-carboline moiety may explain the potent effect of MZA and D hydrochloride salts on PMAstimulated O 2 and TXB 2 . Interestingly, the fused 13-membered macrocyclic amine and octahydroisoquinoline ring system, as well as substitutions in these rings appeared to be a non-factor for the in vitro activity of the manzamines. Finally, the reported pharmacokinetic properties and the lack of significant in vivo toxicity [14] of MZA, a β-carboline alkaloid whose complete synthesis has been reported [11], would suggest that MZA is a prime candidate for further investigation of its potential utility as a pharmacophore from which new and novel therapeutic agents for neuroinflammatory diseases might be developed.

Reagents
LPS B (Escherichia coli 026:B6) was obtained from Difco Laboratories (Detroit, MI); Wright Giemsa stain (modified), ferricytochrome c type III (from horse heart) (FCC), superoxide dismutase (from bovine liver), phorbol 12myristate 13-acetate (PMA), zymosan and dimethyl sulphoxide (DMSO) were obtained from Sigma Chemical Co. (St. Louis, MO). PMA was maintained at -80°C as a 10 mM stock solution in DMSO. Opsonized zymosan (OPZ) was maintained at -20°C in a stock solution of 15 mg/ml in PBS and prepared as described [65]. Dulbecco's modified Eagle medium (DMEM) with high glucose (4,500 mg/l), Hank's balanced salt solution (HBSS), penicillin (P), streptomycin (S), trypsin (0.25%)-EDTA (1 mM) and trypan blue were purchased from GIBCO-BRL (Grand Island, NY); certified heat-inactivated fetal bovine serum (FBS) was obtained from Hyclone (Logan, UT); a LPS stock of 1 mg/ml was prepared in a 0.9% sodium chloride nonpyrogenic solution from Baxter Healthcare Corp. (Toronto, ONT, Canada) and then diluted with DMEM plus 10% FBS plus P and S to the appropriate concentration used in our experiments. Both the LPS stock solution [10 ng/ml] and dilutions were stored at -80°C, thawed prior to each experiment and discarded after use.

LPS containment
To inactivate extraneous LPS, all glassware and metal spatulas were baked for 4 hours at 180°C. Sterile and LPS-free 75-and 162-cm 2 vented cell culture flasks, 24-well flatbottom culture clusters, 96-well cell culture clusters and disposable serological pipettes were purchased from Costar Corporation (Cambridge, MA), while polystyrene cell culture dishes (60 × 15 mm) were obtained from Corning Glass Works (Corning, NY). Sterile and LPS-free Eppendorf Biopur pipette tips were purchased from Brinkmann Instruments, Inc. (Westbury, NY).

Manzamines A, B, C, D, E and F
Manzamine A (MZA) was isolated from a marine sponge species of the genus Haliclona collected off Manzamo, Okinawa in waters at a depth of 30 m in April 1985 [10]. Manzamine B (MZB), manzamine C (MZC) and manzamine D (MZD), were isolated from a sponge of the genus Amphimedon collected by SCUBA off Manzamo, Okinawa [66]. Manzamine E (MZE) and manzamine F (MZF) were isolated from a sponge species of the genus Xestospongia collected off the coast of Miyako Island, Okinawa in June 1986 [67]. All manzamines were dissolved in DMSO to prepare a 10 mM stock and stored at -80°C prior to use in the experiments.

Isolation and culture of rat neonatal microglia
All experiments were performed with adherence to the National Institutes of Health guidelines on the use of experimental animals and with protocols approved by Midwestern University's Research and Animal Care Committee. To isolate rat neonatal microglia, cerebral cortices of 1-2 day-old Sprague-Dawley rats purchased from Harlan (Indianapolis, IN) were surgically removed and placed in cold DMEM + 10% FBS + 120 U/ml P and 12 µg/ml S, the meninges carefully removed, and brain tissue minced and dissociated with trypsin-EDTA at 36°C for 3-5 min. The mixed astroglial cell suspension was plated in either 75-or 162-cm 2 vented cell culture flasks with DMEM medium supplemented with 10% FBS + 120 U/ml P + 12 µg/ml S and grown in a humidified 5% CO 2 incubator at 36°C for 12-14 days. On day 14 and every 3-4 days thereafter, microglia were detached using an orbital shaker (150 rpm, 0.5 hours, 36°C, 5% CO 2 ), centrifuged (400 × g, 25 min, 4°C), and microglia number and viability assessed by trypan blue exclusion. Microglia were characterized as described earlier [23]. Depending on the particular experimental protocol (see below), microglia averaging greater than 95% viability were plated in 24well cell culture clusters, with DMEM supplemented with 10% FBS + 120 U/ml P + 12 µg/ml S, and placed in a humidified 5% CO 2 incubator at 36°C 18-24 hours prior to the experiments.

Experimental protocol to study the effect of manzamines A -F on microglia release of TXB 2 and O 2 -
To study the effects of manzamines A, B, C, D, E and F on the generation of TXB 2 and O 2 -, rat neonatal microglia (2 × 10 5 cells/24-well cell culture clusters) were treated to the following protocol. Seventeen hours prior to the experiments, microglia cells were treated with LPS (0.3 ng/ml) in a final volume of 1 ml of DMEM supplemented with 10% FBS + 120 U/ml P + 12 µg/ml. Thereafter the media was removed and replaced with 1 ml warm HBSS, one of the manzamines (0.1-10 µM final concentration) or vehicle (DMSO) was added, and the microglia incubated for fifteen minutes in a humidified 5% CO 2 incubator at 35.9°C. After the fifteen minute preincubation period with either manzamines or vehicle, PMA (1 µM) or OPZ (0.5 mg/mL) was added and microglia incubated for 70 minutes in a humidified 5% CO 2 incubator at 35.9°C in the presence of the manzamines or vehicle. The final concentration of DMSO did not affect microglia viability or LDH release. O 2 -, TXB 2 and lactate dehydrogenase (LDH) release were assayed as described below.
Assay for TXB 2 Following the incubation of LPS-activated microglia with HBSS, manzamines or vehicle as explained above, PMA-(1 µM) or OPZ-(0.5 mg/mL)triggered TXB 2 generation in the culture supernatants was measured using immunoassays (Cayman Chemical, Ann Arbor, MI) as indicated by the manufacturer's protocol. Results were expressed as pg/mL produced after 70 min of PMA or OPZ stimulation.

Assay for O 2 -
O 2generation was determined by the SOD-inhibitable reduction of FCC [23]. Briefly, PMA (1 µM) or OPZ (0.5 mg/mL)-triggered O 2release from LPS-activated microglia was measured in the presence of FCC (50 µM) and HBSS, with or without SOD (700 Units) which inhibited >95% of FCC reduction, during the 70 min incubation described above. All experimental treatments were run in triplicate and in a final volume of 1 ml. Changes in FCC absorbance were measured at 550 nm using a Beckman DU-650 spectrophotometer. Differences in the amount of reduced FCC in the presence and absence of SOD were used to determine O 2 generation and expressed in nmol by employing the molecular extinction coefficient of 21.0 × 10 3 M -1 cm -1 .

Experimental protocol to study the effect of manzamines on superoxide anion by the hypoxanthine-xanthine oxidase system
A standard hypoxanthine-xanthine oxidase system was used as a cell-free source of O 2 -. O 2was generated by incubation of purified xanthine oxidase (0.02 Units/ml) with hypoxanthine (1.5 mM) at 37°C [21]. O 2 formation was assessed spectrophotometrically as the increase in absorbance at 550 nm associated with the SOD (30 U/mL)inhibitable reduction FCC (50 µM) as described above for rat microglia O 2generation and expressed in nmoles/30 minutes.

Lactate dehydrogenase assay
Lactate dehydrogenase (LDH) release from microglia was determined spectrophotometrically as described elsewhere [9]. Microglia LDH release was expressed as a percentage of total LDH. Total LDH resulted from 0.1% Triton X-100-lysed microglia cells (intracellular LDH) plus LDH released to the extracellular medium.

Statistical analysis of the data
Data were analyzed with the Prism ® software package purchased from GraphPad (San Diego, CA.). One-way analysis of variance followed by Dunnett's test was performed on all sets of data. Manzamine-treated groups were compared with the vehicle-treated group, shown as 0 or control in the corresponding figures. Differences were considered statistically significant at p < 0.05 and reported in each figure legend.